Improving OFI Transport Counter Polling Implementation#1218
Draft
markbrown314 wants to merge 2 commits intoSandia-OpenSHMEM:mainfrom
Draft
Improving OFI Transport Counter Polling Implementation#1218markbrown314 wants to merge 2 commits intoSandia-OpenSHMEM:mainfrom
markbrown314 wants to merge 2 commits intoSandia-OpenSHMEM:mainfrom
Conversation
markbrown314
changed the title
markbrown314
force-pushed
the
1217/development
branch
from
332f14d to
bdc7bbc
Compare
Do not force hard polling when XPMEM is enabled. This conflicts with counter based polling when used with OFI transport. Issue Sandia-OpenSHMEM#1217 Signed-off-by: Mark F. Brown <mark.f.brown@intel.com>
Replaced complex polling with simpler OFI counter wait Issue Sandia-OpenSHMEM#1217 Signed-off-by: Mark F. Brown <mark.f.brown@intel.com>
markbrown314
force-pushed
the
1217/development
branch
2 times, most recently
from
269cb34 to
b096e60
Compare
bcmIntc
reviewed
Feb 3, 2026
| cntr_put_attr.events = FI_CNTR_EVENTS_COMP; | ||
| cntr_get_attr.events = FI_CNTR_EVENTS_COMP; | ||
|
|
||
| #if 0 |
Collaborator
There was a problem hiding this comment.
Did you mean to leave this dead code?
Collaborator
Author
There was a problem hiding this comment.
Yes, see comment below.
|
|
||
| /* wait for put counter to meet outstanding count value */ | ||
|
|
||
| /* Note: the communication routines increment pending put counters before |
Collaborator
There was a problem hiding this comment.
You might want to update this text, as it seems no longer correct.
Collaborator
Author
There was a problem hiding this comment.
This part is incomplete. The original code had this complex counter wait mechanism, before it performs an OFI counter wait. I need to run some more analysis on why it was implemented this way in the first place (Chesterton's fence principle) before removing it completely.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The current libfabric counter polling implementation requires a series of separate lock acquisitions fi_cntr(), fi_cntr_err(), and fi_cq_read(), that can be simplified if we only use fi_cnt_wait(). We see improvements in put latency benchmarks when the shmem_transport_ofi_put_quiet() polling code is removed and no change in functionality.
Addresses Issue #1217
For evaluation purposes only.